Multi-joint active excitation for automatically traversing structural dynamics of milling robot workspace using only output data

The real evaluation of the structural dynamics of a milling robot is essential for real-time monitoring and accurate vibration suppression. However, the multi-joint structure of the robot results in a large number of postures. Current dynamics analysis methods usually select a few postures at stationary state for hammering experiments, which makes it difficult to traverse the dynamics of the robot in the whole workspace, and it is difficult to characterize the dynamics of the robot in the operating state using the dynamics parameters in the stationary state. To achieve a complete vibration analysis, this paper proposes a multi-joint active excitation operational modal analysis. The proposed method meets the input white noise assumption and spatial uniformity requirements by controlling multiple joints simultaneously for random start-stop, uses the resulting random pulse signals excite the overall structure in multiple directions. Then, analyze and simulate the effect of joint control parameters on excitation frequency and energy. Finally, three multi-joint active excitation experiments with different postures are conducted. The results demonstrate that the proposed method can reliably identify the posture-dependent milling robot modal parameters in the operating state, which can be applied to milling robot chatter analysis, path optimization and other real-time process requirements.